Gp2x Joystick Exposed - The Internals


I think this problem has been blown out of all proportion. Kudos to the guys working on a more permanent solution, but even if we are stuck with a slightly unusual stick it isn't the end of the world. If there are proven work-arounds that devs can make use of, we can at least minimise the problem.

Besides, I'm starting to get used to the finicky nature of the stick. It takes a while to become accustomed to the large deadzone after playing with the GP32's super-sensitive stick (which is also a stick I didn't like to begin with), but after a while you hardly even notice it. The diagonal bias is obviously a different problem altogether, although hopefully even that can be lessened with the work-arounds suggested earlier.

I wouldn't waste time trying persuade GPH to ammend the problem through a hardware fix. Eolair is bang on the money - they won't want to change such an integral hardware component after shipping so many units and we'd be stuck with two different hardware versions. Stupid.

A ghetto fix might help solve the problem for individuals with enough technical ability, but for mere mortals like myself we'll just have to live with it. Yet thanks to the considerate devs out there *hint hint ;)* it'll hardly even register on the radar, given time...
 
it hasnt been blown out of proportion at all, anyone who thinks its fine obviously hasnt played any very demanding games on their 2X, the stick SUCKS, i was hoping there would be a way to fix it somehow but it looks like were totally fucked, or is there no way someone can mod the firmware / applications so it works "ok"?
 
it hasnt been blown out of proportion at all, anyone who thinks its fine obviously hasnt played any very demanding games on their 2X, the stick SUCKS, i was hoping there would be a way to fix it somehow but it looks like were totally fucked, or is there no way someone can mod the firmware / applications so it works "ok"?

Of course there is a software solution here. There is nothing wrong with this stick itself except that it's unusual and most games were coded having a generic 4 switch pads in a mind.

Look at the famous now DaveC configuration types for gp2x's stick.

The "case 2" and "case 5" will improve handling significantly and some ported games doesn't have much problems with it already. Such functionality should be included in a firmware but isn't yet. So (imho) gp2x's stick will not be that bad thought it can be now. Yet I'm wondering sometimes why GPH didn't chose a generic 4-way stick instead?
 
Last edited by a moderator:
Yet I'm wondering sometimes why GPH didn't chose a generic 4-way stick instead?

Answer: They thought they were being clever :p

I also think case 2 and 5 will help. Although not completely. You only have to play a couple of games like International Karate+ and Giana Sisters in Frodo to realize that (Frodo uses case 2). So there's still something missing but hopefully that can be fixed with some clever coding.

The best thing I think would be if DaveC's config choices were incorporated into all emus (kinda like miq01's config tester) so that the user could pick the best joystick config for any particular game.
 
Last edited by a moderator:
Ugh, anyone getting that sinking feeling that GPH wasn't the -good- hardware side of the company all of a sudden?

Someone -needs- to contact GPH about these problems.

That would take time and money though and you'll never see it on the GP2X you have in your hand. That, I think it would be even worse because we would end up with two versions that would behave differently and require different software to be playable.

Not entirely. If GPH rotated the stick 45 degrees it would not effect any of the software options. Case 2 would still be the way to go for the same stuff as the current one. All stick options would hold. The only difference is that it would behave a little better than the current ones.

So I would still pull for GPH to change the orientation to 45 deg. Although I think we would have better luck trying to raise the Titanic, it is worth a shot though.

We need to convince Craigix first as he has close contact with GPH. This means he has a .00000001% better chance of not getting ignored by GPH than we do :p
 
Last edited by a moderator:
2 quick things,

1st, I don't quite get the shape of the contactor hat from your picture, could you maybe get some pics of that piece from a different angle?

2nd, if we did rotate it by 45­° would we be able to modify the hat with a shim or something to decrease the deadzone?
 
I've been speaking to them about the stick for weeks...

If they don't want to use a totally different type which would fuck all of us that have them, they could do the easy thing and just rotate the assembly 45 degrees to help with diagonals.

If they went and switched to a 4-way pad or stick all of us that have units would not be able to use software made for the new one and you would have to have different versions for each. We would have the BLU/BLU+ mess all over again, we definetly don't want that.
 
Last edited by a moderator:
Ugh, anyone getting that sinking feeling that GPH wasn't the -good- hardware side of the company all of a sudden?

Someone -needs- to contact GPH about these problems.

That would take time and money though and you'll never see it on the GP2X you have in your hand. That, I think it would be even worse because we would end up with two versions that would behave differently and require different software to be playable.

Not entirely. If GPH rotated the stick 45 degrees it would not effect any of the software options. Case 2 would still be the way to go for the same stuff as the current one. All stick options would hold. The only difference is that it would behave a little better than the current ones.

So I would still pull for GPH to change the orientation to 45 deg. Although I think we would have better luck trying to raise the Titanic, it is worth a shot though.

We need to convince Craigix first as he has close contact with GPH. This means he has a .00000001% better chance of not getting ignored by GPH than we do :p
The problem is that I think you would see a lot of people demanding refunds/exchanges if GPH made an important hardware change like that.

Anyway, I did an update on my joystick figure (config 1). I'm thinking that the empty space between the connectors are actually the intermediate diagonals. If you put the stick in a empty space it will make a logical (AND) connection between two connectors. It's hard to tell though without having an actual disassembled stick to look at. Maybe DaveC can confirm it if it's true.

stick67wr.jpg


As you can see this would mean that my earlier calculations were slightly off (the 34/55 ratio). In my new figure the horizontal/vertical moves only have a action margin of 25° each, diagonals have 45° each and intermediate diagonals have 10° each (making it a 25/45/10 distribution).

Converting this to millimeters of tolerance it would mean that horizontal/vertical moves have a hit area of 1.09mm, diagonals 1,96mm and intermediate diagonals 0,43mm (I've earlier estimated the circumference of the circle of movement to be a total of 15.7mm).

Also using config 1 in miq01's joystick config tester I'd say these numbers are pretty accurate. The intermediate diagonals are almost impossible to hit, the horizontal/verticals are somewhat tricky while the diagonals are easy.

Using case 2 to add the intermediate diagonals to the horizontal/vertical moves we can see that it increases the horizontal/vertical margin to 45° each, while keeping the diagonals at 45° each. Working correctly that would pretty much give us an unbiased joystick. Getting it to work could be pretty tricky though. As both DaveC and me have noted.

Edit: after trying some more with the joystick tester I've noted that it's possible to hit a dead area where the joystick doesn't register any input at all (not to be confused with the dead zone). The dead areas are placed where the intermediate diagonals are. They don't appear if you rotate the stick, even if you do it very slowly. The only way to hit them seems to be if you move directly from an upright position to an intermediate diagonal (it's very difficult to hit them). I'm going to have to ask miq01 to change the program though because it's not visually accurate enough to drawn any conclusions.
 
Last edited by a moderator:
We would have the BLU/BLU+ mess all over again, we definetly don't want that.

Yes, that totally wrecked the community, scars and nightmares are still with us... :(

(It wasn't that big of a deal, and this early there wouldn't even be abandoned project which would never get the fix...)
 
Last edited by a moderator:
The only time I really notice it is when playing a side scroller with up mapped to jump and I will suddenly jump when i hit diagonals accidentally.

I think maybe we could change the button mapping within open2x's sdl code to the 2nd config so that it doesnt require change within people projects. I think if the stick config and the hardware accelleration (+ scaler when supported) is all implemented into open2x's cvs we should then have a pretty consistent version of sdl all round that works.

HH mode on the other hand will have to be left to the devs, but seeing as HH has some of the community's top devs working on it im sure it wont be a problem for them.
 
LOL After reading all, and thinking its all kind of rubbish. It finally becomes clear, that DaveC's original diagram is completely off the mark.

After all the hoohaa Its just a boring 8-way joystick. well thats exacty what is needed for 90% of arcade games(all the modern ones) need :).

I vote to burn DaveC's original diagram.

joystick.png


You have to admire someone who's gone to such lengths to prove themselves wrong

UPDATE: now have this picture as my background so I can laugh all day long :D

Its too late at night to tell people :(
 
The most ideal solutions needs to be in software As not everyone is adept at sticking stuff inside a peice of technology that they have very little knowledge of. Setting up games to play with case 2 for 2D games maybe. You know while they arnt as reliant on the full range of motion, you still have to occasionally shoot diagnally. 3D games could be programed to use the controller in one of the more diagnal friendly variations.

Is this at all feasible? If so it;d be nice to get started on doing just that because playing Tilematch is a pain when half my left, right (as well as Up and down) presses go unregognised because it registeres as a diagnal.
 
Back
Top